US11657454B2 - System and method for constructing optimized ESG investment portfolios - Google Patents

System and method for constructing optimized ESG investment portfolios Download PDF

Info

Publication number
US11657454B2
US11657454B2 US16/419,257 US201916419257A US11657454B2 US 11657454 B2 US11657454 B2 US 11657454B2 US 201916419257 A US201916419257 A US 201916419257A US 11657454 B2 US11657454 B2 US 11657454B2
Authority
US
United States
Prior art keywords
esg
alpha
data
factor
return
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active, expires
Application number
US16/419,257
Other versions
US20190362427A1 (en
Inventor
Mike Chen
George D. Mussalli
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Panagora Asset Management Inc
Original Assignee
Panagora Asset Management Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Panagora Asset Management Inc filed Critical Panagora Asset Management Inc
Priority to US16/419,257 priority Critical patent/US11657454B2/en
Publication of US20190362427A1 publication Critical patent/US20190362427A1/en
Application granted granted Critical
Publication of US11657454B2 publication Critical patent/US11657454B2/en
Active legal-status Critical Current
Adjusted expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/06Asset management; Financial planning or analysis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00Administration; Management
    • G06Q10/06Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/067Enterprise or organisation modelling
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00Handling natural language data
    • G06F40/20Natural language analysis
    • G06F40/279Recognition of textual entities
    • G06F40/284Lexical analysis, e.g. tokenisation or collocates
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00Handling natural language data
    • G06F40/30Semantic analysis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0282Rating or review of business operators or products

Definitions

  • the present invention relates to a system and a method for constructing optimized Environmental, Social and governance (ESG) portfolios, and in particular to a system and a method for constructing investment portfolios that optimize traditional alpha objectives combined with ESG objectives.
  • ESG Environmental, Social and governance
  • FIG. 1 shows the continuum of ESG products 50 offered by asset managers ranging from purely ESG-focused (impact investing) 60 to ESG agnostic (classic investing) 52 .
  • Another consideration is which ESG issues specifically should a portfolio incorporate and address? While there is universal agreement amongst asset owners that outperforming their stated benchmark is desirable, there is much less agreement when it comes to ESG incorporation. For example, one asset owner may primarily be concerned with environmental pollution, while another may care more about gender equality and workplace diversity.
  • the current ESG offerings can be grouped as follows:
  • the present invention relates to a system and a method for constructing investment portfolios that optimize traditional alpha objectives combined with ESG objectives.
  • the invention features a system for constructing an environmental, social and governance (ESG) optimized investment portfolio including an ESG database, an alpha database, a user interface, and an ESG optimization engine.
  • the ESG database includes ESG data for one or more investment products used to construct an investment portfolio.
  • the alpha database includes alpha data for the one or more investment products used to construct the investment portfolio and the alpha data comprise rate of return data for the one or more investment products.
  • the user interface is used for receiving a user's input for selecting an ESG optimization factor.
  • the ESG optimization engine includes an ESG optimization application that receives alpha data for the one or more investment products of the investment portfolio, ESG data for the one or more investment products of the investment portfolio and the user's ESG optimization factor and calculates a combined alpha-ESG rate of return for the investment portfolio, a combined alpha-ESG factor for the investment portfolio and a covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio.
  • the system also includes a computing system that has at least a processor configured to execute computer-implemented instructions of the ESG optimization application.
  • Implementations of this aspect of the invention may include one or more of the following features.
  • R t comprises an alpha rate of return for the investment portfolio for time period t
  • R t ESG comprises any third party, unbiased measure of the user's selected ESG optimization factor for time period t
  • is a tunable scalar
  • product ⁇ R t ESG maps a measure of the selected ESG optimization factor for time period t to a real number and adjusts it by the tunable scalar
  • R t c is the combined rate of return at the end of time period t.
  • the combined alpha-ESG factor for the investment portfolio is calculated via
  • F i,t is an alpha factor
  • F ESG j,t is an ESG optimization factor
  • ⁇ i , ⁇ ESG j are weights
  • F c t is the combined alpha-ESG factor for the investment portfolio at the beginning of time period t.
  • the covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio is used to calculate an excess rate of return during time period t via
  • cov(F c t , R t c ) is the covariance between the combined alpha-ESG factor F c t and combined alpha-ESG rate of return R t c
  • ⁇ (F c t , R t c ) is a correlation of the combined alpha-ESG factor F c t and the combined alpha-ESG rate of return R t c
  • ⁇ (F c t ) is a standard deviation of the combined alpha-ESG factor F c t
  • ⁇ (R t c ) is the standard deviation of the alpha-ESG rate of return R t c .
  • the ESG data comprise publicly available ESG vendor data and PanAgora ESG data.
  • the PanAgora ESG data comprise one of employee sentiment, ESG dictionary, management sentiment, women on board, ESG spin, ESG momentum, and unusual ESG news.
  • the PanAgora ESG data comprise employee sentiment data and the employee sentiment data are collected using web-scraping techniques.
  • the PanAgora ESG data comprise an ESG dictionary that comprises words extracted from sustainability reports.
  • the PanAgora ESG data comprise management sentiment data that are extracted from company internal documents and communications using the ESG dictionary and calculating aggregate textual tone.
  • the PanAgora ESG data comprise ESG spin data and the ESG spin data comprise a difference between publicly available ESG data and ESG management sentiment.
  • the PanAgora ESG data comprise unusual ESG news collected using web-scraping techniques and analyzing the collected data.
  • the publicly available ESG vendor data comprise data from one of OwlShare, TR Asset 4, MSCI ESG, RepRisk, and Sustainalytics, among others.
  • the invention features a method for constructing an environmental, social and governance (ESG) optimized investment portfolio including the following.
  • an ESG optimization engine comprising an ESG optimization application that receives alpha data for the one or more investment products of the investment portfolio, ESG data for the one or more investment products of the investment portfolio and the user's ESG optimization factor and calculates a combined alpha-ESG rate of return for the investment portfolio, a combined alpha-ESG factor for the investment portfolio and a covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio.
  • Providing a computing system comprising at least a processor configured to execute computer-implemented instructions of the ESG optimization application.
  • the invention features a method for constructing an ESG dictionary including the following. Reading all words from a plurality of sustainability reports. Next, removing all stop words and replacing uppercase letters with lowercase letters. Next, ranking the words by frequency and constructing a word frequency list. Next, extracting a group of nouns from the word frequency list. Next, assigning polarity to each noun manually. Next, extracting a group of verbs and a group of adjectives from the word frequency list. Next, adding antonyms of each verb and each adjective to the group of verbs and group of adjectives, respectively. Finally, assigning polarity to each verb and each adjective automatically using a Python built-in dictionary.
  • Polarity is assigned to each word manually by first entering a word into the ESG dictionary, then checking to see if the word is ESG related, then removing the word if it is not ESG related, then assigning polarity to the word if it is ESG related, wherein the polarity comprises one of positive, negative or neutral values.
  • the method further includes calculating an ESG Spin by taking the difference between an externally provided ESG rating and the ESG management sentiment.
  • m i c ⁇ w 1 ,w 2 ,w 3 ,w 4 ⁇ i /c ⁇ w 1 ,w 2 ,w 3 ⁇ I .
  • FIG. 1 depicts the current ESG investing landscape
  • FIG. 2 A depict schematic diagrams of a process for constructing ESG investment portfolios according to this invention
  • FIG. 2 B depicts a schematic diagram of a framework for constructing ESG investment portfolios according to this invention
  • FIG. 3 depicts a flow diagram for the method of constructing ESG investment portfolios, according to this invention.
  • FIG. 4 depicts a flow diagram for the integration process of alpha factors with ESG factors and with the client's input, according to this invention
  • FIG. 5 depicts a list of examples of third party ESG Vendors
  • FIG. 6 depicts a list of examples of internally developed PanAgora specific ESG factors
  • FIG. 7 A depicts employee sentiment data for Nvidia extracted from the web
  • FIG. 7 B is a plot of Nvidia share price over a period of time
  • FIG. 8 is a plot of the employee sentiment data and the stock performance for Netflix over the same time period
  • FIG. 9 A depicts a WordCloud of All Words from 500 sustainability reports used to construct an ESG Dictionary
  • FIG. 9 B depicts a WordCloud of verbs from the 500 sustainability reports used to construct an ESG Dictionary
  • FIG. 9 C depicts a WordCloud of adjectives from the 500 sustainability reports used to construct an ESG Dictionary
  • FIG. 10 depicts a flow diagram for the process of constructing the ESG Dictionary
  • FIG. 11 depicts a flow diagram for the process of assigning polarity to the nouns in the ESG Dictionary
  • FIG. 12 A depicts the rule for assigning polarity to the combinations of positive verb/adjective with a noun
  • FIG. 12 B depicts the rule for assigning polarity to the combinations of negative verb/adjective with a noun
  • FIG. 13 depicts a flow diagram for calculating ESG manager sentiment
  • FIG. 14 is a graph of the average cumulative returns for companies as a function of the time of disclose of negative news
  • FIG. 15 is a flow diagram for the process of calculating the ESG Spin factor
  • FIG. 16 depicts examples of companies with high/low Truth scores
  • FIG. 17 A and FIG. 17 B depict a flow diagram for the process of constructing a measure of the “unusual ESG news factor”
  • FIG. 18 A is a graph of the unusualness factor for the Volkswagen AG case
  • FIG. 18 B is a graph of the negative sentiment score for the Volkswagen AG case
  • FIG. 19 A depicts the process of establishing materiality of ESG alpha factors through contextualization in order to determine which ESG-alpha factors are most relevant to each company;
  • FIG. 19 B depicts ESG-alpha factors and the corresponding sub-universes of the investible universe
  • FIG. 20 depicts the overall investment process using the method of constructing ESG investment portfolios, according to this invention.
  • FIG. 21 depicts the ESG performance of an example of an ESG investment portfolio constructed according to the methodology of this invention
  • FIG. 22 depicts an exemplary computing system architecture 600 for implementing the system of the present invention
  • FIG. 23 is a schematic diagram of an exemplary computer system 500 that is used to implement the system of the present invention.
  • the present invention provides a system and a method for constructing ESG portfolios that optimally combine profit maximizing characteristics with ESG characteristics as well as client-specific requirements.
  • the stock selection ESG-based process takes into consideration alpha signals ( ⁇ ) with long term investment horizons 95 , industry specific signals 94 , and event-driven signals 92 .
  • the stock selection process also takes into consideration the client's input regarding relative importance of ESG performance versus alpha performance ( ⁇ ), and the client's input regarding the ESG metrics that matter the most to them (X).
  • an integrated ESG, alpha and impact tracking framework 80 for constructing ESG-Alpha portfolios includes ESG vendor data 82 , PanAgora specific ESG data 84 , and PanAgora equity alpha data 88 .
  • the ESG vendor data 82 and the PanAgora specific ESG data 84 are combined to generate a combined ESG rating 86 and the combined ESG rating 86 and the equity alpha data 88 are used as inputs for the process of constructing an integrated ESG-Alpha portfolio 90 .
  • a method 100 of constructing ESG investment portfolios that optimize traditional alpha objectives combined with ESG objectives, based on the framework 80 includes the following step. First, we take industry recognized third party ESG Vendor factors 82 and their measures ( 102 ). Next, we take internally developed PanAgora specific ESG factors 84 and their measures ( 104 ). We combine the third party ESG Vendor factors 82 and the internally developed PanAgora specific ESG factors 84 applying PanAgora developed materiality considerations to provide combined ESG factors 86 and their measures ( 106 ). Next, we receive the client's (or investor's) input and selection regarding the ESG optimization criterion ( 108 ).
  • F i,t is an ⁇ factor
  • F ESG j,t is an ESG factor
  • ⁇ i , ⁇ ESG j are weights.
  • Combined return R t c is the return during the period t (available at the end of the period) and combined factor F c t (and its constituent factors) is available at the beginning of period t ( 118 ).
  • the single period excess return ⁇ t is expressed as a function of the covariance between the combined factor F c t and the risk-adjusted combined return R t c :
  • the covariance cov(F c t , R t c ) between the combined factor F c t and the risk-adjusted combined return R t c is written in term of the correlation ⁇ (F c t , R t c ) of the combined factor F c t and the risk-adjusted combined return R t c and the standard deviations ⁇ (F c t ) and ⁇ (R t c ).
  • the risk-aversion parameter is expressed as:
  • ⁇ v c [ v , v ESG ] T
  • ⁇ i , j , t c ⁇ ⁇ ( F i , t c , F j , t c )
  • ⁇ model is the target tracking error, which is proportional to the cross-sectional dispersion of the forecast ⁇ (F c t ) and the square root of the number of stocks N, but inversely proportional to the risk-aversion parameter ⁇ t .
  • the information ratio (IR) that compares the manager's excess returns to the amount of risk is the ratio of the average IC to the standard deviation of the IC:
  • the optimal weights ⁇ * are calculated by taking the derivative of the IR c equation and setting the derivative equal to zero, since this is an unconstrained optimization.
  • IR * IC _ T ⁇ ⁇ IC c , - 1 IC _
  • the time-invariant factor correlation and weight do not really hold in practice.
  • the scalar parameter ⁇ can be used to tune how much do we want to tilt toward ESG factors F ESG i in the portfolio. For example, if we want to split the model evenly between alpha and ESG factors, then we set ⁇ such that:
  • the ESG vendor data 82 are provided by industry recognized ESG vendors. Examples of these ESG vendors include OwlShare, TR Asset4, MSCI ESG, RepRisk, and Sustainalytics, among others, as shown in FIG. 5 .
  • the PanAgora ESG data 84 are additional ESG data that are developed internally, as will be described below. Examples of the PanAgora ESG data 84 include Employee Sentiment 302 , ESG Dictionary 304 , Management ESG Sentiment 306 , Women on Board/C-suite 307 , ESG Spin 308 , ESG Momentum 309 , and Unusual ESG News 310 , as shown in FIG. 6 .
  • Employee sentiment 302 provides an inside view of a company, and can be a leading indicator of the firm's financial performance.
  • Employee sentiment data 302 are collected using web-scraping techniques to “read” millions of direct comments from employees on their thoughts of management, how likely they are to recommend their company to a friend, and their overall job satisfaction.
  • the employee sentiment 302 for Nvidia Company is extracted via web-scraping from websites such as Glassdoor.com.
  • the employee sentiment rating 312 is based on questions about culture and values, work/life balance, senior management, compensation and benefits and career opportunities.
  • a statistical overall average rating 314 is plotted for a given time period.
  • the Nvidia stock performance (A) follows the same upward trend as the employee sentiment rating 302 for the same time period, as shown in FIG. 7 B .
  • the employee sentiment (R) and the share price for Netflix (NFLX) plotted over the same time period are shown in FIG. 8 . It is shown that the split of the streaming business and the DVD by mail business increased the employee sentiment (R) and eventually increased the overall stock performance (B).
  • an internal ESG dictionary 304 is constructed using the following process 320 .
  • the manual process 340 of assigning polarity to ESG nouns includes the following. First, each word (i.e., noun) is entered in the dictionary ( 342 ) and then the word is checked to see if it is ESG related ( 344 ). If the noun is not ESG related, it is removed ( 345 ). If the noun is ESG related, polarity is assigned ( 346 ). Three polarity groups are constructed, i.e., positive ( 347 ), neutral ( 348 ) and negative ( 349 ).
  • Management behavior insights are also used to develop ESG-alpha factors.
  • ESG-alpha factors there is a common behavioral tendency for C-Suite executives to avoid disclosing negative news too early.
  • This behavioral tendency we verified that upon disclosure of negative ESG news, companies tend to work to remedy their ESG related issue, and this results in improved ESG alpha factors.
  • the act of disclosing ESG related issues indicates that ESG is important to a firm's management.
  • Management ESG Sentiment 306 is constructed by developing an internal ESG dictionary 304 , reading through millions of companies' internal corpus to identify and assign relevant ESG information using natural language processing (NLP) techniques, and then applying machine learning techniques to access the relative impact of ESG comments.
  • NLP natural language processing
  • the process for constructing an ESG manager sentiment 360 includes the following.
  • a positive verb/adjective combined with a positive ESG noun is positive
  • a positive verb/adjective combined with a neutral ESG noun is positive
  • a positive verb/adjective combined with a negative ESG noun is negative
  • a negative verb/adjective combined with a positive ESG noun is negative
  • a negative verb/adjective combined with a neutral ESG noun is negative
  • a negative verb/adjective combined with a negative ESG noun is positive.
  • ESG Manager Sentiment (Number of positive hits ⁇ Number of negative hits)/number of total hits
  • ESG Management Sentiment negatively predicts future stock returns. Managers have incentives to immediately reveal good news to investors, but accumulate and withhold bad news up to a certain threshold. When the managers are forced to disclose bad news, things are likely bottomed out and therefore we can bet on stock price reversal after the disclosure of bad news.
  • FIG. 14 the average cumulative returns for companies is plotted as a function of the time of disclose of negative news. The time of disclose of negative news or when the manager ESG sentiment is negative is set to zero point on the horizontal axis. The horizontal axis extends from 50 days before to 50 days after the date when the negative news are disclosed. We observe that the average cumulative returns for companies decrease before the disclosure of the negative news (i.e., time zero) and then increases.
  • the ESG Spin/Truth factor 308 is the difference between the externally provided ESG score (such as MSCI ESG rating) and the ESG manager sentiment.
  • the ESG Spin/Truth factor 308 is a measure of a company's opaqueness on ESG issues. Referring to FIG. 15 , the process 380 of calculating the ESG Spin factor 308 includes the following. We take the normalized third party ESG rating (e.g., MSCI ESG rating) and subtract the normalized ESG manager sentiment ( 382 ). Next, we normalize the result ( 384 ), we flip the sign ( 386 ) and we obtain the ESGSpin/Truth score ( 388 ). An ESG Spin is indicated when the third party ESG rating is low while the ESG manger sentiment is high. ESG Truth is indicated when the third party ESG rating is consistent with the ESG manger sentiment. Examples of companies with high/low Truth scores are shown in FIG. 16 .
  • Increase in unusual ESG news 310 is also a factor that affects future stock performance.
  • increase in the “unusualness” of news coupled with negative sentiment may predict a decrease in company stock price.
  • the terms “emissions tests”, “fraud accusations against” and “to cheat exhaust emissions” occurred more frequently than in previous years.
  • the phrase “negative outlook on” had appeared 688 times always followed by “any”.
  • October 2011 we observed 13 occurrences of the phrase “negative outlook on France”.
  • a measure of the “unusual ESG news factor” 310 is constructed using the process 400 of FIG. 17 A and FIG. 17 B .
  • Google News platform has 4500 sources including both traditional news websites and blogs.
  • Python newspaper package to download relevant full articles ( 404 ).
  • the evaluation datasets include all the news texts for the current month and the training datasets include all the news text for the previous two years.
  • For each evaluation dataset ( 409 ) we extract n-grams and count frequencies (i th n-gram count c i ).
  • An n-gram is a specified sequence of n words.
  • polarity For each i th n-gram extracted from the evaluation text we assign polarity ( 416 ). The polarity of the n-gram is positive if there is at least one positive word and no negative word. The polarity of the n-gram is negative if there is at least one negative word and no positive word.
  • Positive Sentiment Number of positive n -grams/total number of n -grams
  • Negative Sentiment Number of negative n -grams/total number of n -grams
  • Sentiment (Number of positive n -grams ⁇ Number of negative n -grams)/total number of n -grams
  • the unusualness factor is the sum of the cross-entropy H and the Sentiment score ( 420 ).
  • FIG. 19 B depicts ESG-alpha factors and the corresponding sub-universes of the investible universe.
  • environmental factor 1 we classify companies according to their ratio of Plants Property and Equipment (PP&E) to Total Assets (TA), and divide the investible universe into high physical capital universe and low physical capital universe.
  • PP&E Plants Property and Equipment
  • TA Total Assets
  • social factor 1 the division of the investible universe is into high employee productivity universe and low employee productivity universe.
  • governance factor 1 the division of the investible universe is into high financial slack universe and low financial slack universe.
  • the ESG characteristics of the investment portfolios that are constructed using the above mentioned methodology is measured and reported versus the benchmark, using the client's preferred ESG metric R t ESG .
  • the investment portfolios return is also measured and reported versus the stated benchmark (MSCI World).
  • the over/under performance along both alpha and ESG dimensions is quantified, and is reported on a regular and live basis. Large ESG score detractors can be drilled down into specific ESG incidents driving down the score, and can be attributed to individual ESG factor in the model.
  • the overall investment process using the method of constructing ESG investment portfolios, according to this invention is shown in FIG. 20 .
  • the ESG performance of an example of an ESG investment portfolio constructed according to the methodology of this invention is shown in FIG. 21 . It is shown that the Stock Selector ESG portfolio ranks better on the ESG front than both its benchmark (MSCI World Index) and the MSCI World ESG Universal Index, as measured by MSCI ESG.
  • an exemplary system architecture 600 for implementing the system of the present invention includes a computing system 500 that receives inputs from database 610 .
  • Database 610 receives data inputs from the public cloud 602 , market data providers 604 , ESG data providers 606 and risk model providers 608 .
  • the data from the database 610 are entered into the alpha-ESG optimization engine 630 and are used as inputs for the alpha-ESG model and factor application 620 .
  • the alpha-ESG optimization engine 630 operates in the computing system 500 and also receives client specific inputs via a user interface (UI) 556 .
  • UI user interface
  • the alpha-ESG optimization results are displayed on display 640 .
  • an exemplary computer system 500 or network architecture that may be used to implement the system of the present invention includes a processor 520 , first memory 530 , second memory 540 , I/O interface 550 and communications interface 560 . All these computer components are connected via a bus 510 .
  • processors 520 may be used.
  • Processor 520 may be a special-purpose or a general-purpose processor.
  • bus 510 connects the processor 520 to various other components of the computer system 500 .
  • Bus 510 may also connect processor 520 to other components (not shown) such as, sensors, and servomechanisms.
  • Bus 510 may also connect the processor 520 to other computer systems.
  • Processor 520 can receive computer code via the bus 510 .
  • the term “computer code” includes applications, programs, instructions, signals, and/or data, among others.
  • Processor 520 executes the computer code and may further send the computer code via the bus 510 to other computer systems.
  • One or more computer systems 500 may be used to carry out the computer executable instructions of this invention.
  • Computer system 500 may further include one or more memories, such as first memory 530 and second memory 540 .
  • First memory 530 , second memory 540 , or a combination thereof function as a computer usable storage medium to store and/or access computer code.
  • the first memory 530 and second memory 540 may be random access memory (RAM), read-only memory (ROM), a mass storage device, or any combination thereof.
  • RAM random access memory
  • ROM read-only memory
  • mass storage device or any combination thereof.
  • FIG. 20 one embodiment of second memory 540 is a mass storage device 543 .
  • the mass storage device 543 includes storage drive 545 and storage media 547 . Storage media 547 may or may not be removable from the storage drive 545 .
  • Mass storage device 543 may be a Compact Disc Read-Only Memory (“CDROM”), ZIP storage device, tape storage device, magnetic storage device, optical storage device, Micro-Electro-Mechanical Systems (“MEMS”), nanotechnological storage device, floppy storage device, hard disk device, USB drive, among others.
  • Mass storage device 543 may also be program cartridges and cartridge interfaces, removable memory chips (such as an EPROM, or PROM) and associated sockets.
  • the computer system 500 may further include other means for computer code to be loaded into or removed from the computer system 500 , such as the input/output (“I/O”) interface 550 and/or communications interface 560 .
  • the computer system 500 may further include a user interface (UI) 556 designed to receive input from a user for specific parameters, such as user defined ESG factor to be used to optimize the investment portfolio.
  • UI user interface
  • Both the I/O interface 550 and the communications interface 560 and the user interface 556 allow computer code and user input to be transferred between the computer system 500 and external devices including other computer systems. This transfer may be bi-directional or omni-direction to or from the computer system 500 .
  • Computer code and user input transferred by the I/O interface 550 and the communications interface 560 and the UI 556 are typically in the form of signals, which may be electronic, electromagnetic, optical, or other signals capable of being sent and/or received by the interfaces. These signals may be transmitted via a variety of modes including wire or cable, fiber optics, a phone line, a cellular phone link, infrared (“IR”), and radio frequency (“RF”) link, among others.
  • signals may be electronic, electromagnetic, optical, or other signals capable of being sent and/or received by the interfaces. These signals may be transmitted via a variety of modes including wire or cable, fiber optics, a phone line, a cellular phone link, infrared (“IR”), and radio frequency (“RF”) link, among others.
  • IR infrared
  • RF radio frequency
  • the I/O interface 550 may be any connection, wired or wireless, that allows the transfer of computer code.
  • I/O interface 550 includes an analog or digital audio connection, digital video interface (“DVI”), video graphics adapter (“VGA”), musical instrument digital interface (“MIDI”), parallel connection, PS/2 connection, serial connection, universal serial bus connection (“USB”), IEEE1394 connection, PCMCIA slot and card, among others.
  • the I/O interface connects to an I/O unit 555 such as a user interface (UI) 556 , monitor, speaker, printer, touch screen display, among others.
  • Communications interface 560 may also be used to transfer computer code to computer system 500 .
  • Communication interfaces include a modem, network interface (such as an Ethernet card), wired or wireless systems (such as Wi-Fi, Bluetooth, and IR), local area networks, wide area networks, and intranets, among others.
  • the invention is also directed to computer products, otherwise referred to as computer program products, to provide software that includes computer code to the computer system 500 .
  • Processor 520 executes the computer code in order to implement the methods of the present invention.
  • the methods according to the present invention may be implemented using software that includes the computer code that is loaded into the computer system 500 using a memory 530 , 540 such as the mass storage drive 543 , or through an I/O interface 550 , communications interface 560 , user interface UI 556 or any other interface with the computer system 500 .
  • the computer code in conjunction with the computer system 500 may perform any one of, or any combination of, the steps of any of the methods presented herein.
  • the methods according to the present invention may be also performed automatically, or may be invoked by some form of manual intervention.
  • the computer system 500 , or network architecture, of FIG. 20 is provided only for purposes of illustration, such that the present invention is not limited to this specific embodiment.

Abstract

A system for constructing an environmental, social and governance (ESG) optimized investment portfolio includes an ESG database, an alpha database, a user interface, and an ESG optimization engine. The ESG database includes ESG data for one or more investment products used to construct an investment portfolio. The alpha database includes alpha data for the one or more investment products used to construct the investment portfolio and the alpha data comprise rate of return data for the one or more investment products. The user interface is used for receiving a user's input for selecting an ESG optimization factor. The ESG optimization engine includes an ESG optimization application that receives alpha data for the one or more investment products of the investment portfolio, ESG data for the one or more investment products of the investment portfolio and the user's ESG optimization factor and calculates a combined alpha-ESG rate of return for the investment portfolio, a combined alpha-ESG factor for the investment portfolio and a covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio.

Description

CROSS REFERENCE TO RELATED CO-PENDING APPLICATIONS
This application claims the benefit of U.S. provisional application Ser. No. 62/675,262 filed on May 23, 2018 and entitled SYSTEM AND METHOD FOR CONSTRUCTING OPTIMIZED ESG INVESTMENT PORTFOLIOS, which is commonly assigned and the contents of which are expressly incorporated herein by reference.
FIELD OF THE INVENTION
The present invention relates to a system and a method for constructing optimized Environmental, Social and Governance (ESG) portfolios, and in particular to a system and a method for constructing investment portfolios that optimize traditional alpha objectives combined with ESG objectives.
BACKGROUND OF THE INVENTION
Financial investing decisions are usually made based on the financial rate of return of an investment product. This is the typical criterion for classical (or alpha)-type investing. However, there are other criteria that may be also used to decide where to make an investment. One of these other criteria is a company's adherence to environmental, social and governance (ESG) principles. Companies that incorporate these traits into their core philosophies are seemingly more aligned with stakeholder interests and often have better financial performance over the long run than those companies with the weakest ESG records.
Investors are increasingly adopting strategies that combine these new ESG principles with existing methods of identifying attractive investment opportunities. However, this is still an evolving field and there is no conclusive process to best construct ESG portfolios that optimally combine profit maximizing characteristics with ESG characteristics as well as client-specific requirements.
A wide spectrum of ESG investment solutions exists as a result of various approaches being taken to address a growing desire for ESG portfolios. The challenge for the investment community, however, is that asset owners who invest in ESG-centric strategies assign different relative importance between ESG criteria and alpha investing criteria. FIG. 1 shows the continuum of ESG products 50 offered by asset managers ranging from purely ESG-focused (impact investing) 60 to ESG agnostic (classic investing) 52. Another consideration is which ESG issues specifically should a portfolio incorporate and address? While there is universal agreement amongst asset owners that outperforming their stated benchmark is desirable, there is much less agreement when it comes to ESG incorporation. For example, one asset owner may primarily be concerned with environmental pollution, while another may care more about gender equality and workplace diversity.
The current ESG offerings can be grouped as follows:
    • Restriction-list-based: This approach, commonly called SRI (Socially Responsible Investing), is the earliest form of ESG investing. This method excludes companies involved in controversial topics such as tobacco, gambling, cluster munitions, among others, from the investable universe. Portfolios based on this approach typically result in lower achieved alpha versus their non-exclusion-list counterparts. The smaller opportunity set reduces the maximum obtainable return by the manager. Furthermore, for those stocks in the investment universe, no distinctions are made between relatively good or bad ESG companies. Therefore, this is a rudimentary approach to ESG investing that likely will deliver less than optimal ESG and Alpha results.
    • Integration Investing: In this approach, ESG factors are incorporated into stock selection and portfolio construction considerations. A growing body of research suggests that ESG factors can contribute to long-term financial performance either through increasing upside opportunities or minimizing downside risks. As a result, ESG becomes another criterion upon which stocks are evaluated, much like value, quality, among others. This approach is attractive from the perspective that ESG issues are evaluated only along the alpha dimension.
    • Impact Investing: In this approach, investors direct capital towards companies which provide solutions to social and environmental issues and to affect real world outcomes. One key challenge of this approach is the measurability of outcomes. Increasingly, the United Nations Sustainable Development Goals (SDGs) is emerging as the dominant framework around which impact investment results are measured. Popular ESG vendor metrics are also used to measure a portfolio's impact or ESG rating. In this approach, a given portfolio is evaluated along both ESG and alpha dimensions. Given the increased availability of data and computing power, the investment community is increasingly adopting quantitative approaches. However, when it comes to ESG portfolios, even quantitative managers do not optimally integrate their traditional and ESG factors in such a way as to maximize joint ESG and alpha performance according to reasonable metrics. The reason for this is because in the traditional sense, the manager's job is to maximize alpha per unit of risk.
Accordingly, there is a need for systems and methods for constructing ESG portfolios that optimally combine profit maximizing characteristics with ESG characteristics as well as client-specific requirements.
SUMMARY OF THE INVENTION
The present invention relates to a system and a method for constructing investment portfolios that optimize traditional alpha objectives combined with ESG objectives.
In general, in one aspect, the invention features a system for constructing an environmental, social and governance (ESG) optimized investment portfolio including an ESG database, an alpha database, a user interface, and an ESG optimization engine. The ESG database includes ESG data for one or more investment products used to construct an investment portfolio. The alpha database includes alpha data for the one or more investment products used to construct the investment portfolio and the alpha data comprise rate of return data for the one or more investment products. The user interface is used for receiving a user's input for selecting an ESG optimization factor. The ESG optimization engine includes an ESG optimization application that receives alpha data for the one or more investment products of the investment portfolio, ESG data for the one or more investment products of the investment portfolio and the user's ESG optimization factor and calculates a combined alpha-ESG rate of return for the investment portfolio, a combined alpha-ESG factor for the investment portfolio and a covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio. The system also includes a computing system that has at least a processor configured to execute computer-implemented instructions of the ESG optimization application.
Implementations of this aspect of the invention may include one or more of the following features. The combined alpha-ESG rate of return for the investment portfolio is calculated via
R t c=(1−γ)R t +γR t ESG
wherein Rt comprises an alpha rate of return for the investment portfolio for time period t, wherein Rt ESG comprises any third party, unbiased measure of the user's selected ESG optimization factor for time period t, wherein γ is a tunable scalar and wherein product γRt ESG maps a measure of the selected ESG optimization factor for time period t to a real number and adjusts it by the tunable scalar and wherein Rt c is the combined rate of return at the end of time period t. The combined alpha-ESG factor for the investment portfolio is calculated via
F t c = i = 1 M v i F i , t + j = M + 1 N v j ESG F j , t ESG .
wherein Fi,t is an alpha factor, FESG j,t is an ESG optimization factor, νi, νESG j are weights, and Fc t is the combined alpha-ESG factor for the investment portfolio at the beginning of time period t. The covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio is used to calculate an excess rate of return during time period t via
α t = ( N - 1 ) λ t cov ( F t c , R t c ) = ( N - 1 ) λ l ρ ( F t c , R t c ) σ ( F t c ) σ ( R t c )
wherein cov(Fc t, Rt c) is the covariance between the combined alpha-ESG factor Fc t and combined alpha-ESG rate of return Rt c, wherein ρ(Fc t, Rt c) is a correlation of the combined alpha-ESG factor Fc t and the combined alpha-ESG rate of return Rt c and wherein σ(Fc t) is a standard deviation of the combined alpha-ESG factor Fc t and σ(Rt c) is the standard deviation of the alpha-ESG rate of return Rt c. The ESG data comprise publicly available ESG vendor data and PanAgora ESG data. The PanAgora ESG data comprise one of employee sentiment, ESG dictionary, management sentiment, women on board, ESG spin, ESG momentum, and unusual ESG news. The PanAgora ESG data comprise employee sentiment data and the employee sentiment data are collected using web-scraping techniques. The PanAgora ESG data comprise an ESG dictionary that comprises words extracted from sustainability reports. The PanAgora ESG data comprise management sentiment data that are extracted from company internal documents and communications using the ESG dictionary and calculating aggregate textual tone. The PanAgora ESG data comprise ESG spin data and the ESG spin data comprise a difference between publicly available ESG data and ESG management sentiment. The PanAgora ESG data comprise unusual ESG news collected using web-scraping techniques and analyzing the collected data. The publicly available ESG vendor data comprise data from one of OwlShare, TR Asset 4, MSCI ESG, RepRisk, and Sustainalytics, among others.
In general, in another aspect, the invention features a method for constructing an environmental, social and governance (ESG) optimized investment portfolio including the following. Providing an ESG database comprising ESG data for one or more investment products used to construct an investment portfolio. Providing an alpha database comprising alpha data for the one or more investment products used to construct the investment portfolio and wherein the alpha data comprise rate of return data for the one or more investment products. Next, receiving a user's input for selecting an ESG optimization factor via a user interface. Providing an ESG optimization engine comprising an ESG optimization application that receives alpha data for the one or more investment products of the investment portfolio, ESG data for the one or more investment products of the investment portfolio and the user's ESG optimization factor and calculates a combined alpha-ESG rate of return for the investment portfolio, a combined alpha-ESG factor for the investment portfolio and a covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio. Providing a computing system comprising at least a processor configured to execute computer-implemented instructions of the ESG optimization application.
In general, in another aspect, the invention features a method for constructing an ESG dictionary including the following. Reading all words from a plurality of sustainability reports. Next, removing all stop words and replacing uppercase letters with lowercase letters. Next, ranking the words by frequency and constructing a word frequency list. Next, extracting a group of nouns from the word frequency list. Next, assigning polarity to each noun manually. Next, extracting a group of verbs and a group of adjectives from the word frequency list. Next, adding antonyms of each verb and each adjective to the group of verbs and group of adjectives, respectively. Finally, assigning polarity to each verb and each adjective automatically using a Python built-in dictionary. Polarity is assigned to each word manually by first entering a word into the ESG dictionary, then checking to see if the word is ESG related, then removing the word if it is not ESG related, then assigning polarity to the word if it is ESG related, wherein the polarity comprises one of positive, negative or neutral values.
In general, in another aspect, the invention features a method for constructing an ESG management sentiment including the following. First, providing an ESG dictionary and using the ESG dictionary to develop combinations of verbs and/or adjectives with ESG nouns using rules in FIG. 12A and FIG. 12B. Next, searching for the combinations of verbs and/or adjectives with ESG nouns in company internal documents and communications. Next, calculating an ESG management sentiment via:
ESG Manager Sentiment=(Number of positive hits−Number of negative hits)/number of total hits
The method further includes calculating an ESG Spin by taking the difference between an externally provided ESG rating and the ESG management sentiment.
In general, in another aspect, the invention features a method of constructing an unusual ESG news factor including the following. First, web-scraping to extract news titles for a specific company and links on a search engine for a specific time-period. Next, downloading relevant full articles using a Python newspaper package. Next, generating evaluation datasets and training datasets, wherein the evaluation datasets comprise all news texts for a current month period and the training datasets comprise all news texts for two years previous to the current month period. For each evaluation dataset extracting n-grams and counting frequencies, wherein an n-gram comprises a specified sequence of n-words. Next, calculating probability pi of occurrence of an ith n-gram in an evaluation dataset via:
p i =c i /Σc i.
For each extracted ith n-gram calculating model probability mi of occurrence of the ith n-gram in a training dataset via:
m i =c{w 1 ,w 2 ,w 3 ,w 4}i /c{w 1 ,w 2 ,w 3}I.
Next, calculating a cross-entropy H of the model probability mi with respect to the probability pi via:
H=−Σp i log(m i) for (i=1 to I).
Next, assigning polarity to each ith n-gram extracted from the evaluation text. Next, calculating positive sentiment, negative sentiment and overall sentiment via:
Positive Sentiment=Number of positive n-grams/total number of n-grams
Negative Sentiment=Number of negative n-grams/total number of n-grams
Overall Sentiment=(Number of positive n-grams−Number of negative n-grams)/total number of n-grams.
Finally, calculating the unusual ESG news factor by taking the sum of the cross-entropy H and the Overall Sentiment.
BRIEF DESCRIPTION OF THE DRAWINGS
Referring to the figures, wherein like numerals represent like parts throughout the several views:
FIG. 1 depicts the current ESG investing landscape;
FIG. 2A depict schematic diagrams of a process for constructing ESG investment portfolios according to this invention;
FIG. 2B depicts a schematic diagram of a framework for constructing ESG investment portfolios according to this invention
FIG. 3 depicts a flow diagram for the method of constructing ESG investment portfolios, according to this invention;
FIG. 4 depicts a flow diagram for the integration process of alpha factors with ESG factors and with the client's input, according to this invention;
FIG. 5 depicts a list of examples of third party ESG Vendors;
FIG. 6 depicts a list of examples of internally developed PanAgora specific ESG factors;
FIG. 7A depicts employee sentiment data for Nvidia extracted from the web;
FIG. 7B is a plot of Nvidia share price over a period of time;
FIG. 8 is a plot of the employee sentiment data and the stock performance for Netflix over the same time period;
FIG. 9A depicts a WordCloud of All Words from 500 sustainability reports used to construct an ESG Dictionary;
FIG. 9B depicts a WordCloud of verbs from the 500 sustainability reports used to construct an ESG Dictionary;
FIG. 9C depicts a WordCloud of adjectives from the 500 sustainability reports used to construct an ESG Dictionary;
FIG. 10 depicts a flow diagram for the process of constructing the ESG Dictionary;
FIG. 11 depicts a flow diagram for the process of assigning polarity to the nouns in the ESG Dictionary;
FIG. 12A depicts the rule for assigning polarity to the combinations of positive verb/adjective with a noun;
FIG. 12B depicts the rule for assigning polarity to the combinations of negative verb/adjective with a noun;
FIG. 13 depicts a flow diagram for calculating ESG manager sentiment;
FIG. 14 is a graph of the average cumulative returns for companies as a function of the time of disclose of negative news;
FIG. 15 is a flow diagram for the process of calculating the ESG Spin factor;
FIG. 16 depicts examples of companies with high/low Truth scores;
FIG. 17A and FIG. 17B depict a flow diagram for the process of constructing a measure of the “unusual ESG news factor”;
FIG. 18A is a graph of the unusualness factor for the Volkswagen AG case;
FIG. 18B is a graph of the negative sentiment score for the Volkswagen AG case;
FIG. 19A depicts the process of establishing materiality of ESG alpha factors through contextualization in order to determine which ESG-alpha factors are most relevant to each company;
FIG. 19B depicts ESG-alpha factors and the corresponding sub-universes of the investible universe;
FIG. 20 depicts the overall investment process using the method of constructing ESG investment portfolios, according to this invention;
FIG. 21 depicts the ESG performance of an example of an ESG investment portfolio constructed according to the methodology of this invention;
FIG. 22 depicts an exemplary computing system architecture 600 for implementing the system of the present invention;
FIG. 23 is a schematic diagram of an exemplary computer system 500 that is used to implement the system of the present invention.
DETAILED DESCRIPTION OF THE INVENTION
The present invention provides a system and a method for constructing ESG portfolios that optimally combine profit maximizing characteristics with ESG characteristics as well as client-specific requirements.
Referring to FIG. 2A, the stock selection ESG-based process takes into consideration alpha signals (α) with long term investment horizons 95, industry specific signals 94, and event-driven signals 92. The stock selection process also takes into consideration the client's input regarding relative importance of ESG performance versus alpha performance (λ), and the client's input regarding the ESG metrics that matter the most to them (X). The composite score is expressed as follows:
Score=α+λX
Referring to FIG. 2B, an integrated ESG, alpha and impact tracking framework 80 for constructing ESG-Alpha portfolios includes ESG vendor data 82, PanAgora specific ESG data 84, and PanAgora equity alpha data 88. The ESG vendor data 82 and the PanAgora specific ESG data 84 are combined to generate a combined ESG rating 86 and the combined ESG rating 86 and the equity alpha data 88 are used as inputs for the process of constructing an integrated ESG-Alpha portfolio 90.
Referring to FIG. 3 , a method 100 of constructing ESG investment portfolios that optimize traditional alpha objectives combined with ESG objectives, based on the framework 80, includes the following step. First, we take industry recognized third party ESG Vendor factors 82 and their measures (102). Next, we take internally developed PanAgora specific ESG factors 84 and their measures (104). We combine the third party ESG Vendor factors 82 and the internally developed PanAgora specific ESG factors 84 applying PanAgora developed materiality considerations to provide combined ESG factors 86 and their measures (106). Next, we receive the client's (or investor's) input and selection regarding the ESG optimization criterion (108). Next, we receive proven alpha factors and together with the combined ESG factors 86 and with the client's selected optimization criterion and we use them to construct an optimized ESG-Alpha portfolio via and an optimization algorithm (110). The optimization algorithm optimizes the investment portfolio both in the alpha direction and in the selected ESG factor direction, as will be described below. Finally, we track the alpha and ESG performance of the optimized ESG-Alpha portfolio (112).
Referring to FIG. 4 , a flow diagram of the ESG-Alpha integration process 110 includes the following. First, we receive the client's selection of the ESG factor that is to be used to optimize the investment portfolio against (112). Next, we take any third party, unbiased measure of the client selected ESG factor at time t: Rt ESG (114). Next, we map the measure of the selected ESG factor at time t to real number and adjust it by a tunable scalar γ. γRt ESG to put it in the same range as stock returns for a given optimization time horizon (116). The score γRt ESG is the investor's utility along the ESG dimension. Next, we define the combined return:
R t c=(1−γ)R t +γR t ESG
and the combined alpha and ESG factor:
F t c = i = 1 M v i F i , t + j = M + 1 N v j ESG F j , t ESG
where Fi,t is an α factor, FESG j,t is an ESG factor, and νi, νESG j are weights. For ease of notation, we alternatively refer Fi,t and FESG j,t as Fc i,t and νi and νESG j, as νc i, i=1, . . . , M, M+1, . . . , N. Combined return Rt c is the return during the period t (available at the end of the period) and combined factor Fc t (and its constituent factors) is available at the beginning of period t (118).
The single period excess return αt is expressed as a function of the covariance between the combined factor Fc t and the risk-adjusted combined return Rt c:
α t = ( N - 1 ) λ t cov ( F t c , R t c ) = ( N - 1 ) λ t ρ ( F t c , R t c ) σ ( F t c ) σ ( R t c )
The covariance cov(Fc t, Rt c) between the combined factor Fc t and the risk-adjusted combined return Rt c, is written in term of the correlation ρ(Fc t, Rt c) of the combined factor Fc t and the risk-adjusted combined return Rt c and the standard deviations σ(Fc t) and σ(Rt c).
The covariance cov(Fc t, Rt c) is expressed as:
cov ( F t c , R t c ) = cov ( i = 1 M v i F i , t + j = M + 1 M v j ESG F j , t ESG , R t c ) = i = 1 M v i cov ( F i , t , R t c ) + j = M + 1 M v j cov ( F j , t ESG , R t c ) = [ i = 1 M v i IC i , t σ ( F i , t ) + j = M + 1 M v j ESG IC j , t ESG σ ( F j , t ESG ) ] σ ( R l c )
where the information coefficients (IC) are defined as the cross-sectional correlation coefficients:
IC i,t=ρ(F i,t ,R t c)
IC j,t ESG=ρ(F j,t ESG ,R t c).
The risk-aversion parameter is expressed as:
λ t = N - 1 σ ( F l c ) σ model σ ( F t c ) = v cT Φ t c v c where: v c = [ v , v ESG ] T , Φ t C = ( Φ i , j , l c ) i , j = 1 N , Φ i , j , t c = σ ( F i , t c , F j , t c )
and σmodel is the target tracking error, which is proportional to the cross-sectional dispersion of the forecast σ(Fc t) and the square root of the number of stocks N, but inversely proportional to the risk-aversion parameter λt.
Substituting the equation for the risk-aversion parameter λt into the equation for the single period excess return αt we have:
α t = IC t c N - 1 σ model σ ( R t c ) and IC t c = ρ ( F t c , R t c ) = i = 1 N v i c IC i , l c σ ( F i , l c ) v cT Φ t c v c = i = 1 M v i IC i , t σ ( F i , t ) + i = M + 1 N v j ESG IC j , t ESG σ ( F j , c ESG ) v cT Φ t c v c
Assuming time-invariant factor correlation, time-invariant model weight and factors standardized to unit standard deviation σ(Fc i,t)=1 the information coefficient ICc t equation becomes as follows (120):
IC t c = 1 τ ( i = 1 M v i IC i , t + j = M + 1 N v j ESG IC j , t ESG ) τ = v c , T Φ c v c
With the above-mentioned assumptions the average and standard deviation of the composite information coefficient become:
IC c _ = 1 τ ( i = 1 M v i IC i _ + j = M + 1 N v j ESG IC j ESG _ ) = 1 τ v T IC _ σ ( IC c ) = 1 τ v cT IC C v where : IC _ = ( IC 1 _ , , IC M _ , IC M + 1 ESG _ , , IC N ESG _ ) T IC c = ( ρ i , j , IC c ) i , j = 1
and ρ, is a time-series covariance of factors Fc i and Fc j's IC.
The information ratio (IR) that compares the manager's excess returns to the amount of risk is the ratio of the average IC to the standard deviation of the IC:
IR c = IC c _ σ ( IC c ) = v T IC _ v cT IC c v
The optimal weights ν* are calculated by taking the derivative of the IRc equation and setting the derivative equal to zero, since this is an unconstrained optimization.
( IR c ) v = IC _ v T IC c v - ( v T IC _ ) IC v ( v T IC c v ) 3 / 2 v * = s IC C , - 1 IC _
where s is an arbitrary scalar (122).
The optimal IR*, using the optimal weight ν* is:
IR * = IC _ T IC c , - 1 IC _
Some practical considerations include the following. The time-invariant factor correlation and weight do not really hold in practice. For practical implementations, we use a rolling window to estimate the correlation and optimize accordingly. Even through s can be any arbitrary scalar, in practice we choose it so that the weights sum up to 1. The scalar parameter γ can be used to tune how much do we want to tilt toward ESG factors FESG i in the portfolio. For example, if we want to split the model evenly between alpha and ESG factors, then we set γ such that:
i = 1 M v i = j = M + 1 N v j ESG = 0.5
In practice, we use slightly modified combined return:
R t c=(1=η)R t +ηγR t ESG
where γ is still used to scale ESG measure to the same range as stock returns for a given time period, and η is between 0 and 1 and is used to choose the portfolio's ESG tilt. The above mentioned framework is agnostic to the ESG measure used. As long as the measure can be translated to a numeric score, the framework will work. In practice we use numerical ESG measures in the same range as stock returns, for a given return horizon.
The ESG vendor data 82 are provided by industry recognized ESG vendors. Examples of these ESG vendors include OwlShare, TR Asset4, MSCI ESG, RepRisk, and Sustainalytics, among others, as shown in FIG. 5 .
The PanAgora ESG data 84 are additional ESG data that are developed internally, as will be described below. Examples of the PanAgora ESG data 84 include Employee Sentiment 302, ESG Dictionary 304, Management ESG Sentiment 306, Women on Board/C-suite 307, ESG Spin 308, ESG Momentum 309, and Unusual ESG News 310, as shown in FIG. 6 .
Numerous studies have shown that happier employees are more productive, creative, and overall have fewer turnovers. Disgruntled employees on the other hand can make workplace toxic, leading to lower productivity. Employee sentiment 302 provides an inside view of a company, and can be a leading indicator of the firm's financial performance. Employee sentiment data 302 are collected using web-scraping techniques to “read” millions of direct comments from employees on their thoughts of management, how likely they are to recommend their company to a friend, and their overall job satisfaction. Referring to FIG. 7A, the employee sentiment 302 for Nvidia Company is extracted via web-scraping from websites such as Glassdoor.com. The employee sentiment rating 312 is based on questions about culture and values, work/life balance, senior management, compensation and benefits and career opportunities. A statistical overall average rating 314 is plotted for a given time period. The Nvidia stock performance (A) follows the same upward trend as the employee sentiment rating 302 for the same time period, as shown in FIG. 7B. The employee sentiment (R) and the share price for Netflix (NFLX) plotted over the same time period are shown in FIG. 8 . It is shown that the split of the streaming business and the DVD by mail business increased the employee sentiment (R) and eventually increased the overall stock performance (B).
Referring to FIG. 10 , an internal ESG dictionary 304 is constructed using the following process 320. First, we read words from 500 sustainability report covering different countries from 2007 to 2017 (322). Next, we remove all stop words and replace uppercases with lowercases (324). Next, we rank the words by occurrence frequencies. We then extract a group of nouns 316, shown in FIG. 9A (326). Next, we assign polarity to each noun in the group manually using the process of FIG. 11 (328). Next, we extract a group of verbs 317 and a group of adjectives 318, shown in FIG. 9B and FIG. 9C, respectively (330). Next, we add the antonyms of each verb and adjective to the lists, since most of the verbs and adjectives are positive (332). Next, we assign polarity to each verb and adjective automatically, using a Python built-in dictionary (334). Referring to FIG. 11 , the manual process 340 of assigning polarity to ESG nouns includes the following. First, each word (i.e., noun) is entered in the dictionary (342) and then the word is checked to see if it is ESG related (344). If the noun is not ESG related, it is removed (345). If the noun is ESG related, polarity is assigned (346). Three polarity groups are constructed, i.e., positive (347), neutral (348) and negative (349).
Management behavior insights are also used to develop ESG-alpha factors. As an example, there is a common behavioral tendency for C-Suite executives to avoid disclosing negative news too early. As a result of this behavioral tendency, we verified that upon disclosure of negative ESG news, companies tend to work to remedy their ESG related issue, and this results in improved ESG alpha factors. Furthermore, the act of disclosing ESG related issues indicates that ESG is important to a firm's management. Management ESG Sentiment 306 is constructed by developing an internal ESG dictionary 304, reading through millions of companies' internal corpus to identify and assign relevant ESG information using natural language processing (NLP) techniques, and then applying machine learning techniques to access the relative impact of ESG comments.
Referring to FIG. 13 , the process for constructing an ESG manager sentiment 360 includes the following. We use the ESG dictionary (362) and we combine verbs or adjectives with ESG nouns using the rules shown in FIG. 12A and FIG. 12B (364).
According to the rule 350 a in FIG. 12A, a positive verb/adjective combined with a positive ESG noun is positive, a positive verb/adjective combined with a neutral ESG noun is positive and a positive verb/adjective combined with a negative ESG noun is negative. According to the rule 350 b in FIG. 12B, a negative verb/adjective combined with a positive ESG noun is negative, a negative verb/adjective combined with a neutral ESG noun is negative, and a negative verb/adjective combined with a negative ESG noun is positive. Next, we search for the derived combinations in the corpus of internal documents that can be found on AlphaSense platform (366). Examples of these internal documents include internal documents such as 10K, 10Q, event transcripts such as earning calls, company presentations and press releases, among others. Next, we aggregate the number of positive hits and number of negative hits and we calculate the ESG manager sentiment using the following equation (368):
ESG Manager Sentiment=(Number of positive hits−Number of negative hits)/number of total hits
As was mentioned above, ESG Management Sentiment negatively predicts future stock returns. Managers have incentives to immediately reveal good news to investors, but accumulate and withhold bad news up to a certain threshold. When the managers are forced to disclose bad news, things are likely bottomed out and therefore we can bet on stock price reversal after the disclosure of bad news. Referring to FIG. 14 , the average cumulative returns for companies is plotted as a function of the time of disclose of negative news. The time of disclose of negative news or when the manager ESG sentiment is negative is set to zero point on the horizontal axis. The horizontal axis extends from 50 days before to 50 days after the date when the negative news are disclosed. We observe that the average cumulative returns for companies decrease before the disclosure of the negative news (i.e., time zero) and then increases.
The ESG Spin/Truth factor 308 is the difference between the externally provided ESG score (such as MSCI ESG rating) and the ESG manager sentiment. The ESG Spin/Truth factor 308 is a measure of a company's opaqueness on ESG issues. Referring to FIG. 15 , the process 380 of calculating the ESG Spin factor 308 includes the following. We take the normalized third party ESG rating (e.g., MSCI ESG rating) and subtract the normalized ESG manager sentiment (382). Next, we normalize the result (384), we flip the sign (386) and we obtain the ESGSpin/Truth score (388). An ESG Spin is indicated when the third party ESG rating is low while the ESG manger sentiment is high. ESG Truth is indicated when the third party ESG rating is consistent with the ESG manger sentiment. Examples of companies with high/low Truth scores are shown in FIG. 16 .
Increase in unusual ESG news 310 is also a factor that affects future stock performance. In particular, increase in the “unusualness” of news coupled with negative sentiment may predict a decrease in company stock price. In one example, while researching Volkswagen AG in September 2015, the terms “emissions tests”, “fraud allegations against” and “to cheat exhaust emissions” occurred more frequently than in previous years. In another example, for Glasserman and Mamaysky 2017 up until October 2011, the phrase “negative outlook on” had appeared 688 times always followed by “any”. In October 2011 we observed 13 occurrences of the phrase “negative outlook on France”. A measure of the “unusual ESG news factor” 310 is constructed using the process 400 of FIG. 17A and FIG. 17B. We start by web-scraping to extract news titles about a company and links on websites such as Google News platform (402). Google News platform has 4500 sources including both traditional news websites and blogs. Next, we use Python newspaper package to download relevant full articles (404). We do data cleansing (406) and we generate evaluation datasets (407) and training datasets for each holding period (408). For monthly time-periods, the evaluation datasets include all the news texts for the current month and the training datasets include all the news text for the previous two years. For each evaluation dataset (409) we extract n-grams and count frequencies (ith n-gram count ci). An n-gram is a specified sequence of n words. There are n-grams that occur frequently in the current news (i.e., in the evaluation dataset), but rarely in previous news (i.e., training dataset), as is the case for the Volkswagen AG and the Glasserman and Mamaysky examples. We calculate the observed probability of occurrence of the ith n-gram in the evaluation text by using the following equation (410):
p i =c i /Σc i
For each ith n-gram extracted from the evaluation text we calculate the probability of occurrence of the ith n-gram in the training dataset (411) using the following equation (412):
m i =c{w 1 ,w 2 ,w 3 ,w 4}i /c{w 1 ,w 2 ,w 3}i
Next we calculate the cross-entropy of the model probability mi with respect to the observed probability pi using the following equation (414):
H=−Σρ i log(m i) for (i=1 to I)
For each ith n-gram extracted from the evaluation text we assign polarity (416). The polarity of the n-gram is positive if there is at least one positive word and no negative word. The polarity of the n-gram is negative if there is at least one negative word and no positive word. Next, we calculate the positive sentiment, the negative sentiment and the overall sentiment score (418) using the following equations:
Positive Sentiment=Number of positive n-grams/total number of n-grams
Negative Sentiment=Number of negative n-grams/total number of n-grams
Sentiment=(Number of positive n-grams−Number of negative n-grams)/total number of n-grams
The unusualness factor is the sum of the cross-entropy H and the Sentiment score (420).
We used this methodology to calculate the unusualness factor and the negative sentiment score for the above mentioned Volkswagen AG case, as shown in FIG. 18A and FIG. 18B, respectively. The Volkswagen AG emission scandal started on Sep. 18, 2015 when the United States EPA issued a notice of violation of the Clean Air Act to Volkswagen AG. Examples of n-grams include: “government fines and settlement”, “to cheat exhaust emission”, “into fraud allegations against”. In January 2017, Volkswagen AG set to plead guilty and confirmed a $4.3 billion settlement deal over emission scandal. Examples of n-grams include: “pleaded guilty to criminal”, “confirm billion settlement and”.
Materiality of the ESG factors varies across companies. Environmental issues are important to industrial companies, and not as important to professional service companies. While employee satisfaction is important to professional service companies and not as important to utilities. The common approach to identifying materiality is to segment companies by industry. However, we have discovered this is not the most ideal method to measure materiality. We use the concept of Contextual Modeling which identifies the power of factors such as value, quality, and momentum have different efficacy across various risk contexts. We have applied this technology across the suite of ESG-alpha factors to evaluate materiality of ESG-alpha factors through contextualization and determined which ESG-alpha factors are most relevant to each company. This process cuts through industry group and can identify differences across companies within an industry. For example, Netflix and Yum! Brands are both in the consumer discretionary sector however relevant ESG-alpha (and standard alpha) factors are clearly different for the two companies. The salient company characteristics are used to figure out the important ESG factors for each company, as shown in the process 450 of FIG. 19A. First, we identify the salient company operating features for each ESG factor (452) and then divide the investment universe up according to these salient company features (454), and then we measure the efficacy of the ESG factors in the sub-universes (456). For example, for pollution mitigation, which matters for firms with large installed plants and equipment, we classify firms according to their ratio of Plants Property and Equipment (PP&E) to Total Assets (TA), and divide the investible universe into high and low universes. We then test the efficacy of pollution ESG factor on the sub-universes separately to determine its efficacy and subsequently weight in the combined ESG model. FIG. 19B depicts ESG-alpha factors and the corresponding sub-universes of the investible universe. As was described above, for environmental factor 1, we classify companies according to their ratio of Plants Property and Equipment (PP&E) to Total Assets (TA), and divide the investible universe into high physical capital universe and low physical capital universe. For social factor 1, the division of the investible universe is into high employee productivity universe and low employee productivity universe. For governance factor 1, the division of the investible universe is into high financial slack universe and low financial slack universe.
The ESG characteristics of the investment portfolios that are constructed using the above mentioned methodology is measured and reported versus the benchmark, using the client's preferred ESG metric Rt ESG. In addition, the investment portfolios return is also measured and reported versus the stated benchmark (MSCI World). The over/under performance along both alpha and ESG dimensions is quantified, and is reported on a regular and live basis. Large ESG score detractors can be drilled down into specific ESG incidents driving down the score, and can be attributed to individual ESG factor in the model. The overall investment process using the method of constructing ESG investment portfolios, according to this invention is shown in FIG. 20 . The ESG performance of an example of an ESG investment portfolio constructed according to the methodology of this invention is shown in FIG. 21 . It is shown that the Stock Selector ESG portfolio ranks better on the ESG front than both its benchmark (MSCI World Index) and the MSCI World ESG Universal Index, as measured by MSCI ESG.
Referring to FIG. 22 , an exemplary system architecture 600 for implementing the system of the present invention includes a computing system 500 that receives inputs from database 610. Database 610 receives data inputs from the public cloud 602, market data providers 604, ESG data providers 606 and risk model providers 608. The data from the database 610 are entered into the alpha-ESG optimization engine 630 and are used as inputs for the alpha-ESG model and factor application 620. The alpha-ESG optimization engine 630 operates in the computing system 500 and also receives client specific inputs via a user interface (UI) 556. The alpha-ESG optimization results are displayed on display 640.
Referring to FIG. 23 , an exemplary computer system 500 or network architecture that may be used to implement the system of the present invention includes a processor 520, first memory 530, second memory 540, I/O interface 550 and communications interface 560. All these computer components are connected via a bus 510. One or more processors 520 may be used. Processor 520 may be a special-purpose or a general-purpose processor. As shown in FIG. 20 , bus 510 connects the processor 520 to various other components of the computer system 500. Bus 510 may also connect processor 520 to other components (not shown) such as, sensors, and servomechanisms. Bus 510 may also connect the processor 520 to other computer systems. Processor 520 can receive computer code via the bus 510. The term “computer code” includes applications, programs, instructions, signals, and/or data, among others. Processor 520 executes the computer code and may further send the computer code via the bus 510 to other computer systems. One or more computer systems 500 may be used to carry out the computer executable instructions of this invention.
Computer system 500 may further include one or more memories, such as first memory 530 and second memory 540. First memory 530, second memory 540, or a combination thereof function as a computer usable storage medium to store and/or access computer code. The first memory 530 and second memory 540 may be random access memory (RAM), read-only memory (ROM), a mass storage device, or any combination thereof. As shown in FIG. 20 , one embodiment of second memory 540 is a mass storage device 543. The mass storage device 543 includes storage drive 545 and storage media 547. Storage media 547 may or may not be removable from the storage drive 545. Mass storage devices 543 with storage media 547 that are removable, otherwise referred to as removable storage media, allow computer code to be transferred to and/or from the computer system 500. Mass storage device 543 may be a Compact Disc Read-Only Memory (“CDROM”), ZIP storage device, tape storage device, magnetic storage device, optical storage device, Micro-Electro-Mechanical Systems (“MEMS”), nanotechnological storage device, floppy storage device, hard disk device, USB drive, among others. Mass storage device 543 may also be program cartridges and cartridge interfaces, removable memory chips (such as an EPROM, or PROM) and associated sockets.
The computer system 500 may further include other means for computer code to be loaded into or removed from the computer system 500, such as the input/output (“I/O”) interface 550 and/or communications interface 560. The computer system 500 may further include a user interface (UI) 556 designed to receive input from a user for specific parameters, such as user defined ESG factor to be used to optimize the investment portfolio. Both the I/O interface 550 and the communications interface 560 and the user interface 556 allow computer code and user input to be transferred between the computer system 500 and external devices including other computer systems. This transfer may be bi-directional or omni-direction to or from the computer system 500. Computer code and user input transferred by the I/O interface 550 and the communications interface 560 and the UI 556 are typically in the form of signals, which may be electronic, electromagnetic, optical, or other signals capable of being sent and/or received by the interfaces. These signals may be transmitted via a variety of modes including wire or cable, fiber optics, a phone line, a cellular phone link, infrared (“IR”), and radio frequency (“RF”) link, among others.
The I/O interface 550 may be any connection, wired or wireless, that allows the transfer of computer code. In one example, I/O interface 550 includes an analog or digital audio connection, digital video interface (“DVI”), video graphics adapter (“VGA”), musical instrument digital interface (“MIDI”), parallel connection, PS/2 connection, serial connection, universal serial bus connection (“USB”), IEEE1394 connection, PCMCIA slot and card, among others. In certain embodiments the I/O interface connects to an I/O unit 555 such as a user interface (UI) 556, monitor, speaker, printer, touch screen display, among others. Communications interface 560 may also be used to transfer computer code to computer system 500. Communication interfaces include a modem, network interface (such as an Ethernet card), wired or wireless systems (such as Wi-Fi, Bluetooth, and IR), local area networks, wide area networks, and intranets, among others.
The invention is also directed to computer products, otherwise referred to as computer program products, to provide software that includes computer code to the computer system 500. Processor 520 executes the computer code in order to implement the methods of the present invention. In one example, the methods according to the present invention may be implemented using software that includes the computer code that is loaded into the computer system 500 using a memory 530, 540 such as the mass storage drive 543, or through an I/O interface 550, communications interface 560, user interface UI 556 or any other interface with the computer system 500. The computer code in conjunction with the computer system 500 may perform any one of, or any combination of, the steps of any of the methods presented herein. The methods according to the present invention may be also performed automatically, or may be invoked by some form of manual intervention. The computer system 500, or network architecture, of FIG. 20 is provided only for purposes of illustration, such that the present invention is not limited to this specific embodiment.
Several embodiments of the present invention have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the invention. Accordingly, other embodiments are within the scope of the following claims.

Claims (14)

What is claimed is:
1. A system for constructing an environmental, social and governance (ESG) investment portfolio comprising:
a computing system comprising at least a memory storing computer-executable instructions of an ESG application, and a processor coupled to the memory;
an ESG database comprising ESG data for one or more investment products used to construct an investment portfolio, wherein the ESG database is communicatively coupled to the computing system via a network;
an alpha database comprising alpha data for the one or more investment products used to construct the investment portfolio and wherein the alpha data comprise rate of return data for the one or more investment products, and wherein the alpha database is communicatively coupled to the computing system via the network;
a user interface for receiving a user's input for selecting an ESG selection factor at a time t;
wherein the ESG application comprises computer-executable instructions for extracting alpha data for the one or more investment products of the investment portfolio for a time period ending at time t from the alpha database, and upon receiving the user's ESG selection factor via the user interface, extracting ESG data for the user's selected ESG selection factor from the ESG database for the time period ending at time t, and then calculating a combined alpha-ESG rate of return for the investment portfolio, a combined alpha-ESG factor for the investment portfolio and a covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio; and
wherein the processor executes the computer executable instructions of the ESG application that calculate the combined alpha-ESG rate of return for the investment portfolio via:

R t c=(1−γ)R t +γR t ESG
wherein Rt comprises an alpha rate of return for the investment portfolio for the time period ending at time t, wherein Rt ESG comprises any third party, unbiased measure of the user's selected ESG selection factor at time t, wherein γ is a tunable scalar and wherein product γRt ESG maps a measure of the selected ESG selection factor to a real number and adjusts it by the tunable scalar and wherein Rt c is the combined rate of return at the end of the time period ending at time t; and
wherein the ESG data comprise an ESG dictionary that comprises words extracted from sustainability reports by:
reading all words from a plurality of sustainability reports;
ranking the words by frequency and constructing a word frequency list;
extracting a group of nouns from the word frequency list; and
assigning polarity to each noun manually.
2. The system of claim 1, wherein the combined alpha-ESG factor for the investment portfolio is calculated via
F t c = i = 1 M v i F i , t + j = M + 1 N v j ESG F j , t ESG .
wherein Fi,t is an alpha factor, FESG j,t is an ESG selection factor, νi, νESG j are weights, and Fc t is the combined alpha-ESG factor for the investment portfolio.
3. The system of claim 1, wherein the covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio is used to calculate an excess rate of return during the time period ending at time t via
α t = ( N - 1 ) λ t cov ( F t c , R t c ) = ( N - 1 ) λ t ρ ( F t c , R t c ) σ ( F t c ) σ ( R t c )
wherein cov(Fc t, Rt c) is the covariance between the combined alpha-ESG factor Fc t and combined alpha-ESG rate of return Rt c, wherein ρ(Fc t, Rt c) is a correlation of the combined alpha-ESG factor Fc t and the combined alpha-ESG rate of return Rt c and wherein σ(Fc t) is a standard deviation of the combined alpha-ESG factor Fc t and σ(Rt c) is the standard deviation of the alpha-ESG rate of return Rt c.
4. The system of claim 1, wherein the ESG data comprise publicly available ESG vendor data and privately-generated ESG data.
5. The system of claim 4, wherein the privately-generated ESG data comprise one of employee sentiment, the ESG dictionary, management sentiment, number of women on board, ESG spin, ESG momentum, and ESG news.
6. The system of claim 4, wherein the privately-generated ESG data comprise employee sentiment data and the employee sentiment data are collected using web-scraping techniques.
7. The system of claim 4, wherein the privately-generated ESG data comprise management sentiment data that are extracted from company internal documents and communications using the ESG dictionary and calculating aggregate textual tone.
8. The system of claim 4, wherein the privately-generated ESG data comprise ESG spin data and the ESG spin data comprise a difference between publicly available ESG data and ESG management sentiment.
9. The system of claim 4, wherein the privately-generated ESG data comprise unusual ESG news collected using web-scraping techniques.
10. A system for constructing an environmental, social and governance (ESG) investment portfolio comprising:
a computing system comprising at least a memory storing computer-executable instructions of an ESG application, and a processor coupled to the memory;
an ESG database comprising ESG data for one or more investment products used to construct an investment portfolio, wherein the ESG database is communicatively coupled to the computing system via a network;
an alpha database comprising alpha data for the one or more investment products used to construct the investment portfolio and wherein the alpha data comprise rate of return data for the one or more investment products, and wherein the alpha database is communicatively coupled to the computing system via the network;
a user interface for receiving a user's input for selecting an ESG selection factor at a time t;
wherein the ESG application comprises computer-executable instructions for extracting alpha data for the one or more investment products of the investment portfolio for a time period ending at time t from the alpha database, and upon receiving the user's ESG selection factor via the user interface, extracting ESG data for the user's selected ESG selection factor from the ESG database for the time period ending at time t, and then calculating a combined alpha-ESG rate of return for the investment portfolio; and
wherein the ESG data comprise an ESG dictionary that comprises words extracted from sustainability reports by:
reading all words from a plurality of sustainability reports;
ranking the words by frequency and constructing a word frequency list;
extracting a group of nouns from the word frequency list; and
assigning polarity to each noun manually.
11. A method for constructing an environmental, social and governance (ESG) investment portfolio comprising:
providing a computing system comprising at least a memory storing computer-executable instructions of an ESG application, and a processor coupled to the memory;
providing an ESG database comprising ESG data for one or more investment products used to construct an investment portfolio, wherein the ESG database is communicatively coupled to the computing system via a network;
providing an alpha database comprising alpha data for the one or more investment products used to construct the investment portfolio and wherein the alpha data comprise rate of return data for the one or more investment products, wherein the alpha database is communicatively coupled to the computing system via the network;
receiving a user's input for selecting an ESG selection factor via a user interface at a time t;
extracting alpha data for the one or more investment products of the investment portfolio for a time period ending at time t from the alpha database using the ESG application, and upon receiving the user's ESG selection factor via the user interface, extracting ESG data for the user's selected ESG selection factor for the time period ending at time t from the ESG database, and then calculating a combined alpha-ESG rate of return for the investment portfolio, a combined alpha-ESG factor for the investment portfolio and a covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio; and
wherein the processor executes the computer-implemented instructions of the ESG application that calculate the combined alpha-ESG rate of return for the investment portfolio is calculated via

R t c=(1−γ)R t +γR t ESG
wherein Rt comprises an alpha rate of return for the investment portfolio for the time period ending at time t, wherein Rt ESG comprises any third party, unbiased measure of the user's selected ESG selection factor at time t, wherein γ is a tunable scalar and wherein product γRt ESG maps a measure of the selected ESG selection factor to a real number and adjusts it by the tunable scalar and wherein Rt c is the combined rate of return at the end of the time period ending at time t; and
wherein the ESG data comprise an ESG disctionary that comprises words extracted from sustainability reports by:
reading all words from a plurality of sustainability reports;
ranking the words by frequency and constructing a word frequency list;
extracting a group of nouns from the word frequency list; and
assigning polarity to each noun manually.
12. The method of claim 11, wherein the combined alpha-ESG factor for the investment portfolio is calculated via
F t c = i = 1 M v i F i , t + j = M + 1 N v j ESG F j , t ESG .
wherein Fi,t is an alpha factor, FESG j,t is an ESG selection factor, νi, νESG j are weights, and Fc t is the combined alpha-ESG factor for the investment portfolio.
13. The method of claim 11, wherein the covariance between the combined alpha-ESG factor and the combined alpha-ESG rate of return for the investment portfolio is used to calculate an excess rate of return during the time period ending at time t via
α t = ( N - 1 ) λ t cov ( F t c , R t c ) = ( N - 1 ) λ t ρ ( F t c , R t c ) σ ( F t c ) σ ( R t c )
wherein cov(Fc t, Rt c) is the covariance between the combined alpha-ESG factor Fc t and combined alpha-ESG rate of return Rt c, wherein ρ(Fc t, Rt c) is a correlation of the combined alpha-ESG factor Fc t and the combined alpha-ESG rate of return Rt c and wherein σ(Fc t) is a standard deviation of the combined alpha-ESG factor Fc t and σ(Rt c) is the standard deviation of the alpha-ESG rate of return Rt c.
14. The method of claim 11, wherein the ESG data comprise publicly available ESG vendor data and privately-generated ESG data and wherein the privately-generated ESG data comprise one of employee sentiment, the ESG dictionary, management sentiment, number of women on board, ESG spin, ESG momentum, and unusual ESG news.
US16/419,257 2018-05-23 2019-05-22 System and method for constructing optimized ESG investment portfolios Active 2041-09-13 US11657454B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/419,257 US11657454B2 (en) 2018-05-23 2019-05-22 System and method for constructing optimized ESG investment portfolios

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201862675262P 2018-05-23 2018-05-23
US16/419,257 US11657454B2 (en) 2018-05-23 2019-05-22 System and method for constructing optimized ESG investment portfolios

Publications (2)

Publication Number Publication Date
US20190362427A1 US20190362427A1 (en) 2019-11-28
US11657454B2 true US11657454B2 (en) 2023-05-23

Family

ID=68614699

Family Applications (1)

Application Number Title Priority Date Filing Date
US16/419,257 Active 2041-09-13 US11657454B2 (en) 2018-05-23 2019-05-22 System and method for constructing optimized ESG investment portfolios

Country Status (1)

Country Link
US (1) US11657454B2 (en)

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11188983B1 (en) * 2019-08-30 2021-11-30 Morgan Stanley Services Group Inc. Computer systems, methods and user-interfaces for tracking an investor's unique set of social and environmental preferences
US11615473B2 (en) * 2020-03-05 2023-03-28 Noor SHAKFEH Resilience measurement system
NO20230189A1 (en) * 2020-07-27 2023-02-27 Rockefeller & Co Llc Environmental, social, and governance (esg) performance trends
US20220343433A1 (en) * 2020-12-10 2022-10-27 The Dun And Bradstreet Corporation System and method that rank businesses in environmental, social and governance (esg)
US20220237715A1 (en) * 2021-01-25 2022-07-28 Liam John Murray Methods, systems, apparatuses, and devices for facilitating dynamic sustainability mapping of real estate
US11741547B1 (en) * 2021-03-17 2023-08-29 Interactive Brokers Llc Systems and methods for data normalization for use in scoring and balancing a portfolio
US11790385B2 (en) * 2021-04-13 2023-10-17 S&P Global Inc. ESG forecasting
US20220414560A1 (en) * 2021-06-10 2022-12-29 Impact Cubed Limited, a Registered Private Company of the Bailiwick of JERSEY System and Method for Performing Environmental, Social, and Governance (ESG) Rating Across Multiple Asset Classes
WO2022271431A1 (en) * 2021-06-22 2022-12-29 The Dun And Bradstreet Corporation System and method that rank businesses in environmental, social and governance (esg)
US11875163B1 (en) 2023-09-13 2024-01-16 Morgan Stanley Services Group Inc. System and method for generating a user interface for visualization, in combination, environmental, social, and governance ( esg ) and financial analytics preliminary

Citations (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070162365A1 (en) * 2005-07-27 2007-07-12 Weinreb Earl J Securities aid
US7277864B2 (en) 2004-03-03 2007-10-02 Asset4 Sustainability ratings and benchmarking for legal entities
US7424451B1 (en) * 2000-10-10 2008-09-09 International Business Machines Corporation System and method of solving optimization problems using prestored advanced bases
US8200527B1 (en) 2007-04-25 2012-06-12 Convergys Cmg Utah, Inc. Method for prioritizing and presenting recommendations regarding organizaion's customer care capabilities
US20120296845A1 (en) * 2009-12-01 2012-11-22 Andrews Sarah L Methods and systems for generating composite index using social media sourced data and sentiment analysis
US8380604B2 (en) 2002-06-03 2013-02-19 Research Affiliates, Llc System, method and computer program product for using a non-price accounting data based index to determine financial objects to purchase or to sell
US20130103624A1 (en) * 2011-10-20 2013-04-25 Gil Thieberger Method and system for estimating response to token instance of interest
US8612270B2 (en) 2004-06-12 2013-12-17 James K. Hazy System and method to simulate the impact of leadership activity
US8694402B2 (en) 2002-06-03 2014-04-08 Research Affiliates, Llc Using accounting data based indexing to create a low volatility portfolio of financial objects
US8799044B2 (en) 2009-01-23 2014-08-05 Infosys Limited System and method for customizing product lifecycle management process to improve product effectiveness
US20150057996A1 (en) * 2013-08-22 2015-02-26 Ricoh Company, Ltd. Text processing apparatus and text display system
US9070134B2 (en) 2009-08-18 2015-06-30 Csrhub Llc Systems, methods, and media for evaluating companies based on social performance
US20160117774A1 (en) 2014-10-24 2016-04-28 ENSOGO Analytics LLC Environment, social and governance rating system
US20160210707A1 (en) * 2009-01-20 2016-07-21 Domos, Llc Computer-implemented systems and methods for simulating large scale automatic data combinations
US9466084B2 (en) 2013-02-15 2016-10-11 Thomson Reuters Global Resources Environmental, social and corporate governance linked debt instruments
US20180013891A1 (en) * 2014-03-25 2018-01-11 Joseph Charlson System and Method for Automated Call Distribution

Patent Citations (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7424451B1 (en) * 2000-10-10 2008-09-09 International Business Machines Corporation System and method of solving optimization problems using prestored advanced bases
US8380604B2 (en) 2002-06-03 2013-02-19 Research Affiliates, Llc System, method and computer program product for using a non-price accounting data based index to determine financial objects to purchase or to sell
US8694402B2 (en) 2002-06-03 2014-04-08 Research Affiliates, Llc Using accounting data based indexing to create a low volatility portfolio of financial objects
USRE44098E1 (en) 2002-06-03 2013-03-19 Research Affiliates, Llc Using accounting data based indexing to create a portfolio of assets
US7277864B2 (en) 2004-03-03 2007-10-02 Asset4 Sustainability ratings and benchmarking for legal entities
US8255271B2 (en) 2004-03-03 2012-08-28 Thomson Reuters Global Resources Sustainability ratings for legal entities with data inspection
US8612270B2 (en) 2004-06-12 2013-12-17 James K. Hazy System and method to simulate the impact of leadership activity
US20070162365A1 (en) * 2005-07-27 2007-07-12 Weinreb Earl J Securities aid
US8200527B1 (en) 2007-04-25 2012-06-12 Convergys Cmg Utah, Inc. Method for prioritizing and presenting recommendations regarding organizaion's customer care capabilities
US20160210707A1 (en) * 2009-01-20 2016-07-21 Domos, Llc Computer-implemented systems and methods for simulating large scale automatic data combinations
US8799044B2 (en) 2009-01-23 2014-08-05 Infosys Limited System and method for customizing product lifecycle management process to improve product effectiveness
US9070134B2 (en) 2009-08-18 2015-06-30 Csrhub Llc Systems, methods, and media for evaluating companies based on social performance
US20120296845A1 (en) * 2009-12-01 2012-11-22 Andrews Sarah L Methods and systems for generating composite index using social media sourced data and sentiment analysis
US20130103624A1 (en) * 2011-10-20 2013-04-25 Gil Thieberger Method and system for estimating response to token instance of interest
US9466084B2 (en) 2013-02-15 2016-10-11 Thomson Reuters Global Resources Environmental, social and corporate governance linked debt instruments
US20170024822A1 (en) 2013-02-15 2017-01-26 André Chanavat Environmental, social and corporate governance linked debt instruments
US20150057996A1 (en) * 2013-08-22 2015-02-26 Ricoh Company, Ltd. Text processing apparatus and text display system
US20180013891A1 (en) * 2014-03-25 2018-01-11 Joseph Charlson System and Method for Automated Call Distribution
US20160117774A1 (en) 2014-10-24 2016-04-28 ENSOGO Analytics LLC Environment, social and governance rating system

Also Published As

Publication number Publication date
US20190362427A1 (en) 2019-11-28

Similar Documents

Publication Publication Date Title
US11657454B2 (en) System and method for constructing optimized ESG investment portfolios
US11164075B2 (en) Evaluation method and apparatus based on text analysis, and storage medium
US8744946B2 (en) Systems and methods for credit worthiness scoring and loan facilitation
Jerven Poor numbers: how we are misled by African development statistics and what to do about it
AU2023206202A1 (en) Risk identification and risk register generation system and engine
US20130138577A1 (en) Methods and systems for predicting market behavior based on news and sentiment analysis
US8930247B1 (en) System and methods for content-based financial decision making support
Sheldrake The business of influence: Reframing marketing and PR for the digital age
See-To et al. Market sentiment dispersion and its effects on stock return and volatility
US20110106743A1 (en) Method and system to predict a data value
US20220343433A1 (en) System and method that rank businesses in environmental, social and governance (esg)
Lappeman et al. The effect of nWOM firestorms on South African retail banking
Zengul et al. Revealing research themes and trends in 30 Top‐ranking accounting journals: a text‐mining approach
Semenova et al. Social contagion and asset prices: Reddit's self-organised bull runs
Sapkota et al. Fear sells: determinants of fund-raising success in the cross-section of initial coin offerings
CN107844874A (en) Enterprise operation problem analysis system and its method
US20160171608A1 (en) Methods and systems for finding similar funds
CN109934701A (en) A kind of methods of risk assessment and device of equity pledge
Dorman et al. The efficacy of publicly-available retirement planning tools
Mai et al. The role of social media in the apparel retail sector in Vietnam: A study of Facebook fan pages
Nijholt et al. Following fashion: Visible progressiveness and the social construction of firm value
Liu et al. Technology and cryptocurrency valuation
Chang et al. Incorporating soft information from financial news media for management decisions in dynamic business environments
Wu et al. Market reaction to iPhone rumors
US20220114518A1 (en) Computer system and computer implemented method

Legal Events

Date Code Title Description
FEPP Fee payment procedure

Free format text: ENTITY STATUS SET TO UNDISCOUNTED (ORIGINAL EVENT CODE: BIG.); ENTITY STATUS OF PATENT OWNER: SMALL ENTITY

FEPP Fee payment procedure

Free format text: ENTITY STATUS SET TO SMALL (ORIGINAL EVENT CODE: SMAL); ENTITY STATUS OF PATENT OWNER: SMALL ENTITY

STPP Information on status: patent application and granting procedure in general

Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION

STPP Information on status: patent application and granting procedure in general

Free format text: NON FINAL ACTION MAILED

STPP Information on status: patent application and granting procedure in general

Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER

STPP Information on status: patent application and granting procedure in general

Free format text: NON FINAL ACTION MAILED

STPP Information on status: patent application and granting procedure in general

Free format text: FINAL REJECTION MAILED

STCF Information on status: patent grant

Free format text: PATENTED CASE